Dec 2021 – Aug 2024 · 2 yrs 8 mos · California, United States

• We are building the Machine Learning muscle for Reltio's fast-growing AI centric cloud-native Master Data Management (MDM) Platform. My team is using different ML models including large language models (LLM) and traditional ML techniques for core areas in MDM workflows. Projects range from using Generative AI for customer self help to using LLMs for solving matching. We have built the platform needed to gather, transform and use data for training the models and serving the models at large scale.

Mar 2019 – Dec 2021 · 2 yrs 9 mos · Santa Barbara, California Area

• I managed and led the machine learning team for the UCC BU in LogMeIn (now GoTo). My team built and partnered with internal and external teams to integrate machine learning into our GoTo suite of collaboration products.
• My team was behind the
• Virtual background replacement and blurring for GoToMeeting with end to end delivery of ML models (creating datasets, training, and development of model architecture) with optimization of the engine in web endpoints
• Working with teams for development and delivering in production of Speech to text models for offline and real-time transcriptions powering multiple GoTo product features including generating text transcripts from meeting audio, improving transcription accuracy, and producing real-time closed captioning
• NLP models behind transcript analysis, text summaries, and extracting action items for GoToMeeting
• Patented process and ML models for extracting presentation slides from video recordings and generating PDF documents
• Data pipelines and ML models for the Video recommendation engine powering GoToStage
• My team grew to have multiple domain experts (3 Staff engineers for ML, Infrastructure, and application integration) and Machine Learning Engineers (4-5 Senior and T2) with 1-2 interns joining the team during different periods throughout the year. We were able to establish how our deliverables were able to impact product ARR. You can experience these features in GoTo products.
• We filed, and received multiple patents for LogMeIn (Goto) and established ML processes for data, model training, ML operations, and product and service integration.

Apr 2017 – Mar 2019 · 1 yr 11 mos · Santa Barbara, California Area

• I built and managed the first Machine Intelligence team at LogMeIn building the next generation products and features driven by Machine Learning, Natural Language Processing and harnessing our Big Data. Team delivered the following:
• Foundational technology used for providing transcriptions for GoToMeeting, GoToWebinar and GoToStage
• Built the data pipeline, feature store and model training infrastructure for GoToStage video recommendation system
• Built a series of different video recommendation statistical and machine learning models powering the GoToStage video recommendation system
• Grew the team to 5 Machine Learning engineers

Feb 2017 – Mar 2017 · 1 mo · Santa Barbara, California Area

• Part of the team building the next generation GoToMeeting features using Machine Learning and Natural Language Processing techniques

Apr 2016 – Jan 2017 · 9 mos · Santa Barbara, California Area

• Was the team lead for team building the next generation GoToMeeting features using Machine Learning and Natural Language Processing techniques. The work was the foundation for meeting recordings, voice transcriptions and NLP based features found in GoToMeeting, GoToWebinar and GoToStage products.

Feb 2014 – Apr 2016 · 2 yrs 2 mos · Santa Barbara, California Area

• I worked on GoToTraining as a full stack engineer working on multiple feature development for a $20M+ ARR product line.

Aug 2007 – Dec 2013 · 6 yrs 4 mos · Greater Los Angeles Area

• I was a graduate student working with Professor Ramesh Govindan at USC doing research on Data Privacy, Sensor Networks and Distributed System Design.

May 2007 – Aug 2007 · 3 mos · Bengaluru Area, India

• I interned with the Mobility, Systems and Networking group at Microsoft Research India, Bangalore. My work focused on studying the limitations of the localization techniques available using the sensors on a 2007 smartphone. We studied use of GSM cell phone tower signal to perform localization and were able to report results with accuracy within 200m radius. This work was eventually used for the ACM Sensys 2008 publication titled "Nericell: Rich Monitoring of Road and Traffic Conditions using Mobile Smartphones".

Sep 2006 – Apr 2007 · 7 mos · Kanpur Area, India

• I was funded by the prestigious Research I Foundation (endowed by Mr. N R Narayana Murth, Chairman Infosys) to work on designing a low cost structural health monitoring system for remote and on-demand monitoring of railway bridges and tunnels called BriMon. The work has been published in Mobisys 2008 and I presented the talk at Breckenridge Colorado.

Oct 2005 – Aug 2006 · 10 mos · Kanpur Area, India

• My work was with the Rural Network Initiative at IIT Kanpur (http://www.cse.iitk.ac.in/users/braman/dgp.html). We designed a low power remote switching technique of rural networking equipment using a mechanism called Wake-on-WLAN which addressed problem of limited battery energy availability in power strapped rural areas. This method provided on-demand remote power switching of the high power consuming networking equipment using low power sensor nodes as channel monitors and actuators. This work was published at 2006 World Wide Web Conference (WWW 2006). I presented the paper in Scotland, UK.

Jan 2002 – Jan 2005 · 3 yrs

• I worked on the online and offline handwritten signature verification system. I developed the feature extraction engine which extracted both dynamic and static features from human signature. In case of online signature we captured the signature input using a stylus over Simputer (hand held computer). Offline signatures were obtained from scanned images of the signatures. The final verification was performed by learning a user specific weight distribution for different features.

May 2004 – Jul 2004 · 2 mos · Bengaluru Area, India

• Medical records, both in structured and un-structured form, are rich in data. A knowledge base can thus be created using the information contained in these records. In this work we looked into various techniques of machine learning and knowledge discovery useful for extracting such information. The ultimate goal of this system is to device a assistant software which can be used as a help to doctors. My work involved studying the knowledge content of such data and use of data mining techniques for recovering this knowledge.